Bitz SIPN Presentation

http://www.arcus.org/sipn
Our goals are to
Improve sea ice forecasts
Advance the Sea Ice Outlook
Improve sea ice models
Cecilia Bitz
University of
Washington
Sea Ice Outlook and the Prediction Network
June 2015 Outlooks of September 2015 Extent
1981-2010 Average
2013 & 2014
Median of Outlooks
(dynamical)
Sea Ice Outlook Spatial Distributions
Probability of Sea Ice Presence (SIP) by Model in 2014
Extrapolation
Figure by Ed BlanchardWrigglesworth
Sea Ice Outlook Spatial Distributions
SIP in 2015 vs 2014
Forecast Agreement
2015
2014
Figures C. Bitz
−1
−0.5
0
−1
0.5
−0.5
1
Probability of ice > 1
where observations had
none
0
−1
0.5
−0.5
1
0
0.5
Probability of ice > 1
where ice was observed
in September
1
Sea Ice Outlook Spatial Distributions
Probability of Sea Ice Presence in September in 2015
Multi-model mean forecast (colors)
with observed extent (black line)
July/Aug initialization
May/June initialization
Improved spatial
pattern at shorter
lead time?
0
0
0.2
0.4
0.6
0.8
0.2 0.4 0.6 0.8
Figure by Ed BlanchardWrigglesworth & C. Bitz
1
1
0
0
0.2
0.4
0.6
0.8
0.2
0.4
0.6
0.8
11
May/June IC Forecasts: GMAO, NRL, UCL-Belgium, MetOffice
July/August IC Forecasts: SLATER, NRL, MetOffice
Sea Ice Outlook Spatial Distributions
First Ice-Free Day (IFD) in 2015, Forecast Initialized in June
Cullather / NASA
GMAO
Posey/ NRL
Observed
150
200
250
150
200
Julian Day
May 1= 121
June 1=152
July 1=182
Aug 1=213
Sep 1=244
Figure
by Blanchard10/07/14
Wrigglesworth & Bitz
150
200
250
250
Analysis of 2015
sea ice by sea ice model
and atmosphere
Sea Ice Model
Figure by Alek Petty & Francois Massonnet et al
Sea Ice Model
Sea Ice Model
Figure by Alek Petty & Francois Massonnet et al
Lessons from 2015 and Recommendations for the 2016 Outlook
• 
Better initialization of sea ice condition, including incorporating more observational data…
IceBridge, CryoSat-2 platforms, sea ice temperature.
• 
Encourage contributions of local-scale metrics, (maps of sea ice concentration and variables we
refer to as SIP and IFD).
• 
Provide tools for gridding and analysis
• 
Encourage regional outlooks
• 
Encourage more uncertainty information
• 
Provide both raw and bias corrected forecasts
• 
A funded effort for experiments
• 
Understand the difference between forecasted and observed sea ice concentrations
• 
Evaluate statistical methods using common metrics
•  Create a more systematic approach to forecasting evaluation, with year round
forecasts and using metrics that also evaluate the local-scale.
Extended SIPN
Proposed to Build and Support a Data Portal
Year round forecasts for 0-5 month range
Daily fields for a dozen variables
Public access to
data products
1
5
Data input
Database
Scientists server-side
data processing for
research
4
2
3
Public access
to forecast
visualizations
Automated forecast
data processing and
creation of products and
visualizations